Folding Algorithms for Protein Structure Prediction

Phil Bradley

University of Washington

February 23, 4:15pm


To reach their biologically active state, newly synthesized proteins
spontaneously fold from an extended, linear conformation into a compact
three-dimensional structure.

This remarkable self-assembly process, protein folding, is guided by the
amino acid sequence of the protein to a unique final state. Despite
several decades of intensive study the process by which sequence
determines structure is still not well understood; in particular it is not
currently possible to predict a protein's native three-dimensional
structure given only its sequence.  Recently, however, a class of
prediction algorithms based on protein fragment assembly have made
considerable headway towards the goal of generating low-resolution
structure predictions. In this talk I will introduce the protein folding
problem, describe these new algorithms, highlight their strengths and
weaknesses, and discuss current research directed at improving the
reliability and accuracy of their predictions.


Return to Applied Math Colloquium home page